-
Notifications
You must be signed in to change notification settings - Fork 46.9k
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Update tracked value after resetting radio group #27394
Conversation
The commented assertInputTrackingIsClean fails.
Comparing: 2807d78...26daa54 Critical size changesIncludes critical production bundles, as well as any change greater than 2%:
Significant size changesIncludes any change greater than 0.2%: (No significant changes) |
This also fixes #19860 😮 : https://codesandbox.io/s/react-18-radio-cannot-be-read-only-forked-vrh6v2?file=/src/App.js I don't have much context about this type of code but the fact that it fixes the original issue and another one, is a pretty strong signal for me. |
This is unconditionally set: I don't think we do for |
Hmm it does seem like there is probably a hydration bug too based on reports at vercel/next.js#49499. |
The isHydrating check is suspicious. I don't think that's quite right. If we do need to set it e.g. to set the dirty flag or update the input value tracking, then we'd have to do that for hydration too. The reason we don't want to do it for hydration is so that it doesn't reset. But maybe what we need to do is just |
I think I'll update this to call updateValueIfChanged() on all the inputs after calling updateInput() on all of them. Will also rename "assertInputTrackingIsClean" to not use the word "Clean" since that could seem related to "Dirty" but it's not. In a separate PR I'll plan to add a test for hydration. My expectation is if you interact with radios before the page hydrates then onChange will not fire for that but it will work normally following that (meaning the dirty flag needs to be set and value tracking needs to be up to date). |
I kind of think onChange should fire after the fact on inputs after hydration if it changed. At least for controlled so that it can enter the correct state. We had the same issue that onLoad for img should probably fire if it loaded before hydration. Maybe for 19? We currently don't really have a commit effect for hydration on DOM nodes which limits stuff like this. |
Fixes whatever part of #26876 and vercel/next.js#49499 that #27394 didn't fix, probably. From manual tests I believe this behavior brings us back to parity with latest stable release (18.2.0). It's awkward that we keep the user's state even for controlled inputs. Previously the .defaultChecked assignment done in updateInput() was changing the actual checkedness because the dirty flag wasn't getting set, meaning that hydrating could change which radio button is checked, even in the absence of user interaction! Now we go back to always detaching again.
Fixes whatever part of #26876 and vercel/next.js#49499 that #27394 didn't fix, probably. From manual tests I believe this behavior brings us back to parity with latest stable release (18.2.0). It's awkward that we keep the user's state even for controlled inputs. Previously the .defaultChecked assignment done in updateInput() was changing the actual checkedness because the dirty flag wasn't getting set, meaning that hydrating could change which radio button is checked, even in the absence of user interaction! Now we go back to always detaching again.
Fixes whatever part of #26876 and vercel/next.js#49499 that #27394 didn't fix, probably. From manual tests I believe this behavior brings us back to parity with latest stable release (18.2.0). It's awkward that we keep the user's state even for controlled inputs. Previously the .defaultChecked assignment done in updateInput() was changing the actual checkedness because the dirty flag wasn't getting set, meaning that hydrating could change which radio button is checked, even in the absence of user interaction! Now we go back to always detaching again.
Fixes #26876, I think. Review each commit separately (all assertions pass in main already, except the last assertInputTrackingIsClean in "should control radio buttons"). I'm actually a little confused on two things here: * All the isCheckedDirty assertions are true. But I don't think we set .checked unconditionally? So how does this happen? * #26876 (comment) claims that d962f35...1f248bd contains the faulty change, but it doesn't appear to change the restoration logic that I've touched here. (One difference outside restoration is that updateProperties did previously set `.checked` when `nextProp !== lastProp` whereas the new logic in updateInput is to set it when `node.checked !== !!checked`.) But it seems to me like we need this call here anyway, and if it fixes it then it fixes it? I think technically speaking we probably should do all the updateInput() calls and then all the updateValueIfChanged() calls—in particular I think if clicking A changed the checked radio button from B to C then the code as I have it would be incorrect, but that also seems unlikely so idk whether to care. cc @zhengjitf @Luk-z who did some investigation on the original issue DiffTrain build for [3c27178](3c27178)
### React upstream changes - facebook/react#27417 - facebook/react#27408 - facebook/react#27409 - facebook/react#27405 - facebook/react#27375 - facebook/react#27407 - facebook/react#27365 - facebook/react#27399 - facebook/react#27395 - facebook/react#27394 - facebook/react#27397 - facebook/react#26992 - facebook/react#27388 - facebook/react#27373 - facebook/react#27332
Sorry for late response. @sophiebits didn't know if this PR is anymore the right place to discuss about (also saw that #26876 was closed ) |
@Luk-z Thank you for your diligence. I have now posted a fix in #27443 and verified that it fixes this repro case: https://codesandbox.io/s/react-canary-radio-buttons-latest-not-working-forked-t2wwyc |
@sophiebits I confirm that it works now, thanks! |
Fixes whatever part of #26876 and vercel/next.js#49499 that #27394 didn't fix, probably. From manual tests I believe this behavior brings us back to parity with latest stable release (18.2.0). It's awkward that we keep the user's state even for controlled inputs, so the DOM is out of sync with React state. Previously the .defaultChecked assignment done in updateInput() was changing the actual checkedness because the dirty flag wasn't getting set, meaning that hydrating could change which radio button is checked, even in the absence of user interaction! Now we go back to always detaching again.
Fixes whatever part of #26876 and vercel/next.js#49499 that #27394 didn't fix, probably. From manual tests I believe this behavior brings us back to parity with latest stable release (18.2.0). It's awkward that we keep the user's state even for controlled inputs, so the DOM is out of sync with React state. Previously the .defaultChecked assignment done in updateInput() was changing the actual checkedness because the dirty flag wasn't getting set, meaning that hydrating could change which radio button is checked, even in the absence of user interaction! Now we go back to always detaching again. DiffTrain build for [db69f95](db69f95)
Fixes whatever part of facebook#26876 and vercel/next.js#49499 that facebook#27394 didn't fix, probably. From manual tests I believe this behavior brings us back to parity with latest stable release (18.2.0). It's awkward that we keep the user's state even for controlled inputs, so the DOM is out of sync with React state. Previously the .defaultChecked assignment done in updateInput() was changing the actual checkedness because the dirty flag wasn't getting set, meaning that hydrating could change which radio button is checked, even in the absence of user interaction! Now we go back to always detaching again.
Fixes whatever part of facebook/react#26876 and vercel/next.js#49499 that facebook/react#27394 didn't fix, probably. From manual tests I believe this behavior brings us back to parity with latest stable release (18.2.0). It's awkward that we keep the user's state even for controlled inputs, so the DOM is out of sync with React state. Previously the .defaultChecked assignment done in updateInput() was changing the actual checkedness because the dirty flag wasn't getting set, meaning that hydrating could change which radio button is checked, even in the absence of user interaction! Now we go back to always detaching again. DiffTrain build for [db69f95e4876ec3c24117f58d55cbb4f315b9fa7](facebook/react@db69f95)
Fixes whatever part of facebook/react#26876 and vercel/next.js#49499 that facebook/react#27394 didn't fix, probably. From manual tests I believe this behavior brings us back to parity with latest stable release (18.2.0). It's awkward that we keep the user's state even for controlled inputs. Previously the .defaultChecked assignment done in updateInput() was changing the actual checkedness because the dirty flag wasn't getting set, meaning that hydrating could change which radio button is checked, even in the absence of user interaction! Now we go back to always detaching again.
Fixes facebook#26876, I think. Review each commit separately (all assertions pass in main already, except the last assertInputTrackingIsClean in "should control radio buttons"). I'm actually a little confused on two things here: * All the isCheckedDirty assertions are true. But I don't think we set .checked unconditionally? So how does this happen? * facebook#26876 (comment) claims that facebook/react@d962f35...1f248bd contains the faulty change, but it doesn't appear to change the restoration logic that I've touched here. (One difference outside restoration is that updateProperties did previously set `.checked` when `nextProp !== lastProp` whereas the new logic in updateInput is to set it when `node.checked !== !!checked`.) But it seems to me like we need this call here anyway, and if it fixes it then it fixes it? I think technically speaking we probably should do all the updateInput() calls and then all the updateValueIfChanged() calls—in particular I think if clicking A changed the checked radio button from B to C then the code as I have it would be incorrect, but that also seems unlikely so idk whether to care. cc @zhengjitf @Luk-z who did some investigation on the original issue
Fixes whatever part of facebook#26876 and vercel/next.js#49499 that facebook#27394 didn't fix, probably. From manual tests I believe this behavior brings us back to parity with latest stable release (18.2.0). It's awkward that we keep the user's state even for controlled inputs, so the DOM is out of sync with React state. Previously the .defaultChecked assignment done in updateInput() was changing the actual checkedness because the dirty flag wasn't getting set, meaning that hydrating could change which radio button is checked, even in the absence of user interaction! Now we go back to always detaching again.
Fixes #26876, I think. Review each commit separately (all assertions pass in main already, except the last assertInputTrackingIsClean in "should control radio buttons"). I'm actually a little confused on two things here: * All the isCheckedDirty assertions are true. But I don't think we set .checked unconditionally? So how does this happen? * #26876 (comment) claims that d962f35...1f248bd contains the faulty change, but it doesn't appear to change the restoration logic that I've touched here. (One difference outside restoration is that updateProperties did previously set `.checked` when `nextProp !== lastProp` whereas the new logic in updateInput is to set it when `node.checked !== !!checked`.) But it seems to me like we need this call here anyway, and if it fixes it then it fixes it? I think technically speaking we probably should do all the updateInput() calls and then all the updateValueIfChanged() calls—in particular I think if clicking A changed the checked radio button from B to C then the code as I have it would be incorrect, but that also seems unlikely so idk whether to care. cc @zhengjitf @Luk-z who did some investigation on the original issue DiffTrain build for commit 3c27178.
Fixes #26876, I think. Review each commit separately (all assertions pass in main already, except the last assertInputTrackingIsClean in "should control radio buttons").
I'm actually a little confused on two things here:
.checked
whennextProp !== lastProp
whereas the new logic in updateInput is to set it whennode.checked !== !!checked
.)But it seems to me like we need this call here anyway, and if it fixes it then it fixes it? I think technically speaking we probably should do all the updateInput() calls and then all the updateValueIfChanged() calls—in particular I think if clicking A changed the checked radio button from B to C then the code as I have it would be incorrect, but that also seems unlikely so idk whether to care.
cc @zhengjitf @Luk-z who did some investigation on the original issue